BIOMEDICAL ENGINEERINGAPPLICATIONS, BASIS & COMMUNICATIONS

181

DESIGNING LOW POWER OF SIGMA DELTA MODULATOR FOR BIOMEDICAL APPLICATION

HO-YIN LEE, CHEN-MING HSU, SHENG-CHIA HUANG, YI-WEI SHIH, CHING-HSING LUO
Department of Electrical Engineering, National Cheng Kung University, Tainan, Taiwan

ABSTRACT
This paper discusses the design of micro power Sigma-delta modulator with oversampling technology. This Sigma-delta modulator design is paid special attention to its low power application of portable electronic system in digitizing biomedical signals such as Electro-cardiogram (ECG), Electroencephalogram (EEG) etc. [1]. A high performance, low power second order Sigma-delta modulator is more useful in analog signal acquisition system. Using Sigma-delta modulator can reduce the power consumption and cost in the whole system. The original biomedical signal can be reconstructed by simply applying the digital bit stream from the modulator output through a low-pass filter. The loop filter of this modulator has been implemented by using switch capacitor (SC) integrators and using simple circuitry consists of OpAmps, Comparator and DAC. In general, the resolution of modulator is about 10 bits for biomedical application. In this two order Sigma-delta modulator simulation results of the 1.8V sigma delta modulator show a 68 dB signal-to-noise-anddistortion ratio (SNDR) in 4 kHz biomedical signal bandwidth and a sampling frequency equal to 1 MHz in the 0.18 m CMOS technology. The power consumption is 400 W. It is very suitable for low power application of biomedical instrument design. Biomed Eng Appl Basis Comm, 2005(August); 17: 181-185. Keywords: Sigma-delta modulator, SDM, low power, biomedical application.

1. INTRODUCTION
The application of portable electronic systems such as wireless communication devices, consumer electronics and battery powered biomedical devices increases the requirement for low-voltage and lowpower circuit techniques [2]. Designing of low-voltage circuit can reduce the number of battery cells for low weight and small system size. At the same time, lowpower circuit design can increase the operation time for biomedical application [3]. Sigma-delta modulator has become a usual technique for analog-to-digital (A/D) conversion. This is because sigma-delta modulator circuits are Received: Dec. 8, 2004; Accepted: July 7, 2005 Correspondence: Ching-Hsing Luo, Professor Department of Electrical Engineering, National Cheng Kung University, Tainan, Taiwan E-mail: robinluo@mail.ncku.edu.tw

structured simply with low-accuracy analog parts and very suitable in low frequency, high performance and low power application. The single-bit signals of sigmadelta modulator are converted into multi-bit signals at the Nyquist sampling rate for biomedical application. Thus, currently usable types of biomedical systems with sigma-delta based A/D converter need single-bit conversion hardware including decimation filter. Sigma-delta modulator utilizes a negative feedback loop consist of an integrator, comparator and one-bit DAC that are very simply components. The input analog signal is first integrated and compared with ground using comparator. Its output drives a one bit DAC which switches reference voltages to the summing node of the integrator, minimizing the difference signal. In this paper, the second order sigma delta modulator can be widely utilized for low-power with biomedical applications. Further power savings can be achieved by using relatively low oversampling ratios. The Biomedical signal acquisition system must extract

23

182

Vol. 17 No. 4 August 2005

and amplify the low level signal (0-20mV). High resolution of sigma delta converters allows the small signals to be acquired directly, without the use of instrumentation amplifiers. The large common mode signals can be digitized without saturation and the small differential signals can be recovered with resolutions comparable to conventional methods.

Fig 1. Block diagram of second-order Sigma Delta Modulator.

2. BACKGROUND
Suppressing the power consumption will be widely utilized for biomedical application. At the same time, in order to get accurate data, the simple Sigmadelta modulation is first-selected to incorporate on the system chip definitely. Sigma-delta modulator has proven to be very suitable in low frequency, highperformance application. A simple block diagram of the second order Sigma Delta Modulator is shown in Figure 1, where x (t) and y (t) are the input and the output signal of sigma delta modulator respectively. The values of y (t) have only two levels as the output because of 1-bit ADC which is comparator is simply used. The 1-bit ADC and DAC are both driven by the same clock signal. Due to the non-linearity incorporated in the ADC, simpler model is used for analysis. The system can be viewed in its discrete time in Figure 2. The comparator is replaced by an adder and sum of noise source Q[z] emulating the quantization noise. And the integrator is replaced by the function block of z-1/ (1-z-1). After some mathematical manipulations in the Z-domain, we can get (1) where Q (z), X (z), Y (z) are the Z-transforms of q[n], x[n],and y[n] respectively. We can see that the digital output Y (z) consist of two parts. The first one is the biomedical signal x[n], X (z) and the other one is error signal q[n], Q (z). The error signal has high pass frequency characteristics in the transfer function of (1-z -1)2. Therefore, the original signal can be obtained from the output signal of Sigma-delta modulator by passing though a decimation filter to return to the original signal.

Fig 2. Simplified model for second order Sigma Delta Modulator. circuit implementation of sigma delta ADCs use SC circuits because it are easily simulated, compatible with VLSI CMOS process and insensitive to clock jitter [5]. Second order sigma delta modulator consists of two integrator, 1-bit ADC, and 1-bit DAC. The power of the first integrator is the major consumption to the overall power dissipation in sigma delta modulator. For that reason, a substantial amount of power can be greatly saved by a suitable circuit design. The finite DC-gain, noise and distortion of the operation amplifier will be determined on the first integrator that will reduce the performance of the whole sigma delta modulator, since these errors will be added directly to the input signal. Furthermore, fully differential structure is used to design integrator and shown in Figure 3. The advantages for this design can improve the noise in the sampling time to increase the resolution of device, reduce the even harmonic and increase the input range. At the same time, the correlated double sampling (CDS) is addition to this integrator that can suppress flicker noise, thermal noise and offset voltage. The schematic of a CDS integrator is shown in Figure 3 [5]. It is clear that the SNR can be increased by increasing the oversampling ratio in any order modulator. If oversampling ratio increases, the modulator will be operated in higher speed and consumed more power. So, further power saving can be achieved by using relatively low oversampling ratios (OSR=fs/2fin)

3. PRINCIPLE AND ANALOG CIRCUIT DESIGN

First of all, we must choose between a switchedcapacitor (discrete-time) and a conventional active RC (continuous-time) [4] for integrator in designing any sigma delta modulator. Generally, most integrated

3.1 Operation Amplifier (OpAmp) and Common Mode Feedback (CMFB)

The integrators include operation amplifier (OpAmp) that is the most critical element and mainly influence the whole performance of modulator. In

24

BIOMEDICAL ENGINEERINGAPPLICATIONS, BASIS & COMMUNICATIONS

183

general, the operation amplifier (OpAmp) in the integrator determines on the operation speed and performance of the modulator. It consumes the power proportional to the gain. Approximately 60dB of dc gain is required to prevent performance degradation and maintain a 12-bit dynamic range [5]. The specifications of OpAmp shown in Table I In order to combine the high gain and low power requirement, a fully differential folded-cascode operation amplifier has been used and shown in Figure 4. The gain bandwidth (GBW) and the open-loop gain (Av) of the amplifier are given by (2)

delta and the bit-stream signal can be utilized in the communication application [6]. Thus, power and chip area can be saved. 3.3 1-bit DAC The purpose of a feedback DAC in a closed-loop sigma delta modulator is to convert the modulator s digital output back into an analog form to be compared to its analog input. In essence, the DAC controls the mapping between the analog and digital domains. Therefore, the performance of the closed-loop sigma delta modulator is completely dependent on the accuracy of its feedback DAC and the way in which its output is compared to the modulator input [5].

(3) At the same time, the Common Mode Feedback (CMFB) of an amplifier is a critical in discrete time (DT) for low voltage and low power application, because it will not reduce the output swing and also require little power only. The common mode output level of the amplifier is maintained by the feedback circuitry also shown in Figure 4. This CMFB circuit has a drawback that common mode level depends on the parameter of device, but the advantage of this circuit is current-path-reduction. In the low power design, this is a very important consideration. Anyway, in a well-design circuit, the loop gain must be relatively high. The closed-loop gain is approximately equal to 1/ , where represents the feedback factor and the open-loop gain of the common mode feedback are given by

Fig 3. The circuit schematic of integrator with CDS in the Sigma-Delta Modulator.

(4)

Fig 4. Fully differential folded-cascode operation amplifier. Table : Specifications of OpAmp in the integrator

(5)

3.2 Comparator
The purpose of the comparator in a sigma delta modulator is to quantize a signal in the loop and provide the digital output of the modulator. The performance of the sigma delta modulator is relatively insensitive to offset and hysteresis in the comparator as the two level quantizer, because the effects of these impairments are attenuated in the baseband by second order noise shaping that can be filtered by decimation filter. The regenerative latch shown in Figure 5 has been used to implement in the comparator because periodic answers are required in all loops of sigma

25

184

Vol. 17 No. 4 August 2005

Fig 5. Comparator implemented with the latch. One of the features in the sigma delta modulators is that they can use low-resolution of DACs in their feedback paths but that is not the same meaning as low accuracy. Actuality, if a multibit DAC is used, the linearity of the DAC must match completely the overall linearity requirements of the modulator. For that reason, 1-bit DACs are typically used in sigma delta modulator. Since they only have two levels, all step sizes are identical. There are, however, other mechanisms that can be separated into two categories for nonidealities in the 1-bit DAC are nonidealities associated with the voltage reference and with the way the charge is taken from the reference [7].

Fig 6. The power spectrum density of the 2nd-order Sigma-delta modulator.

4. SIMULATION RESULT
The simulation results of this 2 nd-order Sigmadelta modulator have been carried out for a sampling frequency of 1MHz and a fixed oversampling ratio (OSR) of 128 with a signal is bandwidth 4 kHz. Furthermore, the input signal is sinusoidal and the reference voltages are 1.1V and 0.7V. The simulated power spectrum density of the modulator is shown in Figure 6. Input biomedical signal frequency is 4 kHz and amplitude is equal to 200mV. The final simulation of the modulator may be considering the signal-tonoise-and- distortion ratio (SNDR) because the limit of the SNDR is due to an imbalance in the design which increases the distortion at higher input signal amplitude. The measured performance of the Sigmadelta modulator is summarized in Table II. The power consumption includes the two OpAmp, the comparator, the D/A and the all bias circuits. In general, a specification of actual ADCs usually used in place of the SNR or SNDR is effective number of bits (ENOB), which is an indication of ADC accuracy at a specific input frequency and sampling rate. The ENOB is defined in equation (3) where DR is the dynamic range of the modulator expressed in dB and each bit of resolution is equal to about 6 dB of dynamic range. (3)

Fig 7. Chip layout of the 2 nd order sigma delta modulator. For small biomedical signal levels, distortion is not important implying that the SNR and SNDR are approximately equal. As the input signal level increase, distortion reduces the modulator performance, and the SNDR will be less than the SNR.

5. CONCLUSION
Designing low-power of two order discrete-time sigma delta modulator has been presented. Due to the lower bandwidth of biomedical signals, sigma delta modulation is feasible. A low power sigma delta can be design by simple circuit that either modulating or

26

BIOMEDICAL ENGINEERINGAPPLICATIONS, BASIS & COMMUNICATIONS

185

Table : Specifications of this second order sigmadelta modulator

4 . Samid L, Manoli Y: Micro Power Continuous-Time Sigma Delta Modulator. Conference on European Solid-State Circuits 2003; 165-168 5 . Norsworthy SR, Schreier R, and Temes GC: Deltasigma converters:/ theory, design and simulation. In: IEEE Press, New York, 1997; 223-234. 6 . Fujisaka H, Kurata R, Sakamoto M, Morisue M: Bit-stream signal processing and its application to communication systems. IEE Circuits, Devices and Systems 2002; 149: 159-166. 7 . Razavi B:Design of analog CMOS integrated circuit. In: McGraw-Hill, New York, 2001. 8 . Thompson H, Hufford M, Evans W and Naviasky E: A low-voltage low-power sigma-delta modulator with improved performance in overload condition. IEEE Custom Integrated Circuits Conference, 2004; 519 - 522.

recovering circuit with oversampling technique. Using oversampling technique, not only the restraints in analogy circuit can be relaxed, the quantization noise can also be reduced. After all, the analog biomedical signal can be reconstructed from the digital bit stream of modulator output by simple low pass filter. This modulator has been designed in a 0.18 m CMOS technology and is shown in Figure 7.

ACKNOWLEDGEMENTS
The authors would like to thank partial financial supports from MOE Program for Promoting Academic Excellence of Universities under the grant number EX91-E-FA09-5-4 and National Chip Implementation Center. Authors also thank Center of Micro/Nano technology, and Center of Bioscience & Technology, National Cheng Kung University.

REFERENCE
1. Joseph JC and John MB: Introduction to Biomedical Equipment Technology. In: Prentice Hall. Upper Saddle River, 2001. 2 . Guessab S, Benabes P and Kielbasa R: A passive delta-sigma modulator for low-power applications. IEEE Circuits and Systems 2004; 3: 295-298 3 . Leung SW and Zhang YT: Digitization of electrocardiogram (ECG) signals using delta-sigma modulation. IEEE Engineering in Medicine and Biology Society 1998; 4: 1964 -1966.

27